meka.filters.multilabel

Class SuperNodeFilter

java.lang.Object

weka.filters.Filter

weka.filters.SimpleFilter

weka.filters.SimpleBatchFilter

meka.filters.multilabel.SuperNodeFilter

All Implemented Interfaces:

java.io.Serializable, weka.core.CapabilitiesHandler, weka.core.OptionHandler, weka.core.RevisionHandler

public class SuperNodeFilter
extends weka.filters.SimpleBatchFilter

SuperNodeFilter.java - Super Class Filter. Input:
Data with label attributes, e.g., [0,1,2,3,4]
A desired partition of indices, e.g., [[1,3],[4],[0,2]], filter
Output:
New data with label attributes: [1+3,4,0+2]
(the values each attribute can take are pruned if necessary)

Version:

June 2012

Author:

Jesse Read

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
SuperNodeFilter()

Method Summary

Methods

Modifier and Type	Method and Description
static int	decodeClass(java.lang.String a) ("c_3",'_') -> 3
static int[]	decodeClasses(java.lang.String a) ("c_3+1") -> [3,1]
static java.lang.String[]	decodeValue(java.lang.String a) "C+A+B" -> ["C","A","B"]
weka.core.Instances	determineOutputFormat(weka.core.Instances D)
static java.lang.String	encodeClass(int j) (3,'_') -> "c_3"
static java.lang.String	encodeClass(int[] c_) ([3,1]) -> "c_3+1"
static java.lang.String	encodeClass(java.lang.String c_j, java.lang.String c_k) (["c_3","c_1"]) -> "c_3+1"
static java.lang.String	encodeValue(weka.core.Instance x, int[] indices) (3,1,2) -> "3+1+2"
static java.lang.String	encodeValue(java.lang.String v_j, java.lang.String v_k) (3,1) -> "3+1"
static java.util.HashMap<java.lang.String,java.lang.Integer>	getCounts(weka.core.Instances D, int[] indices, int p) Return a set of all the combinations of attributes at 'indices' in 'D', pruned by 'p'; AND THEIR COUNTS, e.g., {(00:3),(01:8),(11:3))}.
static java.lang.String[]	getNeighbours(java.lang.String y, java.util.ArrayList<java.lang.String> S, int n) GetNeighbours - return from set S, label-vectors closest to y, having no more different than 'n' bits different.
static java.lang.String[]	getNeighbours(java.lang.String y, java.util.HashMap<java.lang.String,java.lang.Integer> C, int n) GetNeighbours - return from set S (the keySet of HashMap C), label-vectors closest to y, having no more different than 'n' bits different.
int	getP()
weka.core.Instance	getTemplate()
static java.util.Set<java.lang.String>	getValues(weka.core.Instances D, int[] indices, int p) Return a set of all the combinations of attributes at 'indices' in 'D', pruned by 'p'; e.g., {00,01,11}.
java.lang.String	globalInfo()
static void	main(java.lang.String[] argv)
static weka.core.Instances	mergeLabels(weka.core.Instances D, int[][] indices, int p, int n) Merge Labels - Make a new 'D', with labels made into superlabels, according to partition 'indices', and pruning values 'p' and 'n'.
static weka.core.Instances	mergeLabels(weka.core.Instances D, int j, int k, int p) Merge Labels.
weka.core.Instances	process(weka.core.Instances D)
void	setIndices(int[][] n)
void	setN(int n)
void	setP(int p)

Methods inherited from class weka.filters.SimpleBatchFilter

allowAccessToFullInputFormat, batchFinished, input

Methods inherited from class weka.filters.SimpleFilter

debugTipText, doNotCheckCapabilitiesTipText, getCapabilities, getDebug, getDoNotCheckCapabilities, getOptions, listOptions, setDebug, setDoNotCheckCapabilities, setInputFormat, setOptions

Methods inherited from class weka.filters.Filter

batchFilterFile, filterFile, getCapabilities, getOutputFormat, getRevision, isFirstBatchDone, isNewBatch, isOutputFormatDefined, makeCopies, makeCopy, mayRemoveInstanceAfterFirstBatchDone, numPendingOutput, output, outputPeek, runFilter, toString, useFilter, wekaStaticWrapper

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

SuperNodeFilter

public SuperNodeFilter()

Method Detail

setIndices

public void setIndices(int[][] n)

setP

public void setP(int p)

getP

public int getP()

setN

public void setN(int n)

determineOutputFormat

public weka.core.Instances determineOutputFormat(weka.core.Instances D)
                                          throws java.lang.Exception

Specified by:

determineOutputFormat in class weka.filters.SimpleFilter

Throws:

java.lang.Exception

getTemplate

public weka.core.Instance getTemplate()

process

public weka.core.Instances process(weka.core.Instances D)
                            throws java.lang.Exception

Specified by:

process in class weka.filters.SimpleFilter

Throws:

java.lang.Exception

encodeClass

public static java.lang.String encodeClass(int j)

(3,'_') -> "c_3"

decodeClass

public static int decodeClass(java.lang.String a)

("c_3",'_') -> 3

encodeClass

public static java.lang.String encodeClass(java.lang.String c_j,
                           java.lang.String c_k)

(["c_3","c_1"]) -> "c_3+1"

encodeClass

public static java.lang.String encodeClass(int[] c_)

([3,1]) -> "c_3+1"

decodeClasses

public static int[] decodeClasses(java.lang.String a)

("c_3+1") -> [3,1]

encodeValue

public static java.lang.String encodeValue(java.lang.String v_j,
                           java.lang.String v_k)

(3,1) -> "3+1"

encodeValue

public static java.lang.String encodeValue(weka.core.Instance x,
                           int[] indices)

(3,1,2) -> "3+1+2"

decodeValue

public static java.lang.String[] decodeValue(java.lang.String a)

"C+A+B" -> ["C","A","B"]

getValues

public static java.util.Set<java.lang.String> getValues(weka.core.Instances D,
                                        int[] indices,
                                        int p)

Return a set of all the combinations of attributes at 'indices' in 'D', pruned by 'p'; e.g., {00,01,11}.

getCounts

public static java.util.HashMap<java.lang.String,java.lang.Integer> getCounts(weka.core.Instances D,
                                                              int[] indices,
                                                              int p)

Return a set of all the combinations of attributes at 'indices' in 'D', pruned by 'p'; AND THEIR COUNTS, e.g., {(00:3),(01:8),(11:3))}.

mergeLabels

public static weka.core.Instances mergeLabels(weka.core.Instances D,
                              int[][] indices,
                              int p,
                              int n)

Merge Labels - Make a new 'D', with labels made into superlabels, according to partition 'indices', and pruning values 'p' and 'n'.

Parameters:

D - assume attributes in D labeled by original index

Returns:

Instances with attributes at j and k moved to position L as (j,k), with classIndex = L-1

mergeLabels

public static weka.core.Instances mergeLabels(weka.core.Instances D,
                              int j,
                              int k,
                              int p)

Merge Labels.

Parameters:

j - index 1 (assume that j < k)

k - index 2 (assume that j < k)

D - iInstances, with attributes in labeled by original index

Returns:

Instaces with attributes at j and k moved to position L as (j,k), with classIndex = L-1

getNeighbours

public static java.lang.String[] getNeighbours(java.lang.String y,
                               java.util.ArrayList<java.lang.String> S,
                               int n)

GetNeighbours - return from set S, label-vectors closest to y, having no more different than 'n' bits different.

getNeighbours

public static java.lang.String[] getNeighbours(java.lang.String y,
                               java.util.HashMap<java.lang.String,java.lang.Integer> C,
                               int n)

GetNeighbours - return from set S (the keySet of HashMap C), label-vectors closest to y, having no more different than 'n' bits different.

globalInfo

public java.lang.String globalInfo()

Specified by:

globalInfo in class weka.filters.SimpleFilter

main

public static void main(java.lang.String[] argv)